Nurse perceptions of organizational culture and its association with the culture of error reporting: a case of public sector hospitals in Pakistan

The aim of this paper was to investigate the association between the organizational culture and the culture of error reporting in the public health care sector, as perceived by nurses. It is agreed that policy improvements in the public sector health care services are possible only when there is adequate empirical research about the status quo [21]. Therefore, the findings of this study will attempt to map a plan for improved organizational culture for nurses, and, consequently, facilitate to improve error reporting and patient safety. The study results are expected to be relevant not only for nurses and other medical practitioners working in the public sector, but also for the private health care sector. Our hypotheses for the study were: (1) When organizational culture is favorable, the error reporting culture will be favorable, and (2) When each of the six subscales of organizational culture are favorable (1. nurse manager ability, leadership and support, 2. nurse participation in hospital affairs, 3. nurse participation in governance, 4. nurse foundations of quality care, 5. nurse coworker communication, 6. nurse staff and resource adequacy), the error reporting culture will also be favorable.

The study was performed in the city of Lahore, which is the capital of the Punjab province in Pakistan. Lahore is the second largest city of Pakistan and the second highest populated city of the country. It is estimated to have more than ten million inhabitants. Data from the website of Pakistan Institute of Medical Sciences official website was used to conveniently sample two geographically spaced tertiary care public sector hospitals from Lahore, out of a total of nine [24]. The two sampled hospitals have been named Hospital A and Hospital B. Both hospitals have high patient turnovers and large in-patient capacities, and are catering to a different set of patients from the rural and urban Lahore District and also from the surrounding villages of Lahore City. Combined, the two hospitals have a large daily out-patient turnover rate of more than 3,800 patients and an in-patient capacity of approximately 1,890 beds.

All registered female nurses who had been working in the hospital for more than one year were sampled. Each designation was sampled, including nurse supervisors, nurse ward heads, nurse instructors, staff nurses and nurse students. Registered nurse students were included in the sample, because they actively perform clinical duties after the first three months of their enrollment as students, and thus they also have rich experience about the state of organizational culture and error reporting in the hospital setting. In Hospital A there is 1 nurse supervisor, 650 staff nurses, 150 nurse ward heads, 20 nurse instructors, and 415 nurse students. In Hospital B (Sheikh Zayed hospital) there is 1 nurse supervisor, 600 staff nurses, 100 nurse ward heads, 13 nurse instructors, and 320 nurse students. Both hospitals combined have a total of 2,270 nurses, with 2 nurse supervisors, 250 nurse ward heads, 33 nurse instructors, 1,250 staff nurses, and 735 nurse students. Yamane’s formulae [25] was used to determine a sample size from an estimated female Punjab nurse workforce of 11,065. Surveys were distributed to the different nurse designations according to their respective weightage. In this way, 35 % of each nurse designation was sampled, including both nurses supervisors, 440 staff nurses, 90 nurse ward heads, 12 nurse instructors, and 260 nurse students.

Survey questions and guidelines were conducted and written in the English language, which is the official academic and working language of the country. Two standardized instruments were used including the “Practice Environment Scale-Nurse Work Index Revised” (PES-NWI) [26], and the “Survey to Solicit Information about the Culture of Reporting” (SSICR) [27]. The validity and reliability of both the PES-NWI [28, 29] and the SSICR [30, 31] has been established by previous research in health care policy improvements. Respondents were provided, through an extensive literature review, with a summarized list of errors that may occur during health care service delivery by medical practitioners (Table 9 in Appendix) [32‐34]. A pretest of the questionnaire was conducted with 35 nurse respondents to ascertain any loopholes that could be rectified before the final administration.

Section I of the questionnaire contained 18 questions pertaining to the socio-demographic characteristics of nurses. These questions include age (‘20–29 years’, ‘30–39 years’ and ‘40+ years’), marital status (‘Never married’, ‘Currently married’ and ‘Divorced/separated/widowed’), regional belonging (‘Punjab’, ‘Sindh’, ‘Baluchistan’ and ‘Khyber Pakhtun Khwan’), religion (‘Muslim’, ‘Christian’, ‘Hindu’ and ‘Ahmedi’), total number of children (‘None’, ‘1–2 children’ and ‘3 or more children’), total monthly income (PKR ‘5,000–19,999’, ‘20,000–39,999’ and ‘≥40,000’), place of residence (‘college hostel’, ‘hospital resident colony’ and ‘private home’), highest nursing degree (‘Nursing Diploma’, ‘BSc Nursing’ and ‘MSc Nursing’), current designation (‘Nurse supervisor’, ‘Nurse instructor’, ‘Nurse ward head’, ‘Staff nurse’ and ‘Student nurse’), type of labor contract (‘Permanent’ and ‘Contractual’), employment status (‘Full-time’ and ‘Part-time’), current government grade (‘16 grade’ and ‘17 grade’), additional employment in private sector (‘Yes’ and ‘No’), and additional hours worked at the public sector hospital during the night, day or evening (‘Yes’ and ‘No’).

Section II contained 31 questions from the PES-NWI, which measures the organizational culture of a hospital by a composite score which is aggregated to the unit level [26]. Items on the scale consist of 5 subscales which measure different aspects of organizational culture, including: (1) nurse manager ability, leadership and support (5 items), (2) nurse participation in hospital affairs (7 items), (3) nurse participation in governance (2 items), (4) nurse foundations of quality care (10 items), (5) nurse-coworkers relations (3 items), and (6) nurse staffing and resource adequacy (4 items).

Section III contained seven questions from the SSICR. The questions measure the culture of error reporting in a hospital and indicate how comfortable the respondent is in sharing errors at the workplace with supervisors and coworkers.

Both the scales have a 4 point rating scale, which indicate respondent extent of agreement with each item. The response categories include: 1 = strongly agree, 2 = agree, 3 = disagree, 4 = strongly disagree. The instrument is scored by calculating mean subscale scores and a total composite score for each respondent which can range from 1–4. Higher scores indicate a less favorable organizational culture and culture of error reporting. Two of the seven items from the error reporting variables had to be reverse coded so that the scoring was aligned across all items of the tool. Reliability analysis was conducted for both the PES-NWI and the SSICR to confirm Cronbach’s alphas of scales. A Cronbach’s alpha of above 0.7 is considered a reliable measure for health and social science research [35]. For this study, the overall internal consistency ranged satisfactorily between values of 0.743 to 0.881.

The questionnaire was distributed at the two hospitals in the time period from November 2013 to January 2014. All nurse employees are required to sign an attendance register daily, placed in the offices of their respective nurse ward head, before the start of their shift. Hence all nurse designations visited the nurse ward head office daily at the start of three different shifts of either 08:00 am, 02:00 pm or 08:00 pm. It was deemed suitable to communicate with nurses at this place of contact to recruit interested participants. Nurses were asked to read the cover letter attached to the survey and select a time to complete the survey in a reserved room of the nursing school of each respective hospital. Surveys were filled in nursing school class rooms, specifically reserved for data collection, with 15–30 nurses at a time. The classrooms afforded privacy and a comfortable setting, away from the hospital building, clinical wards, patients and attendants, work pressures, male coworkers and other work-related intrusions. The survey completion time fell between 20–35 min. The first author was present to answer questions related to the survey. Respondents sealed their completed surveys and dropped them in a box before leaving the nursing school. Both the nurse supervisors from each hospital were sampled, 440 surveys were distributed to staff nurses, 90 to nurse ward heads, 12 to nurse instructors and 260 to nurse students (Table 1). In total, 804 questionnaires were distributed, but only 309 nurses pre-booked a time to complete the survey in reserved rooms. All 309 nurses completed the survey and were included in the final analysis (response rate: 34.8 %).

Raw data was first entered into Excel. It was then transferred into SPSS 17.0 for analysis. A significance level of 0.05 was assigned for all statistical analyses. First, a factor analysis was used to reduce data and confirm subscales of relevance for the study. Factor analysis was deemed suitable to validate the NES-PWI for Pakistani public sector hospitals [36‐38], as this tool, to the best of researchers’ knowledge, has not been used in the region before. Principal component analysis (PCA) was used with varimax rotation, as guided by previous research [26, 39, 40]. The following conditions for PCA were met: (i) sample size of above 50 cases, (ii) normal distribution, and (iii) all the variables of organizational culture correlating with each other above 0.2 and no correlations of above 0.9 (avoiding fears of multicollinearity).

Descriptive statistics are provided to show socio-demographic and employment characteristics of nurse respondents. Composite scores for subscales of organizational culture and error reporting were calculated [41‐45]. Mean scores were also calculated ranging from 1–4 for subscales of organizational culture, composite organizational culture and error reporting. As recommended by literature, scores under the values of 2.5 were considered favorable and scores of above the value of 2.5 were considered unfavorable. Simple bivariate logistic regression and multivariable logistic regression models were used [46‐49] to check for the association between: (i) organizational culture and its subscales with error reporting, and (ii) nurse socio-demographic characteristics and error reporting. The aim was to identify the odds of a favorable error reporting culture when organizational culture and its subscales are favorable, and also to identify the odds of a favorable error reporting culture in relation to the socio-demographic features of the nurse. The enter method was used. Variables were recoded into bivariate categories in order to use logistic regression. Organizational culture and its six subscales were recoded with dummy variables of 0 = unfavorable organizational culture and 1 = favorable organizational culture. Error reporting was recoded with a dummy variable of 0 = unfavorable error reporting culture and 1 = favorable error reporting culture. The significance of the main effects was estimated by computing the confidence level for Exp (B) and was presented in form of odds ratios (OR), with accompanying 95 % confidence intervals (95 % CI). Each of the variables of organizational culture, its subscales and the socio-demographic characteristics of nurses were adjusted for nurse age (as a continuous variable), nurse literacy and nurse monthly income.

Ethics committee permission was obtained from the Institutional Review Board, University of the Punjab, and also from the hospitals and nursing institutes where data collection took place. The ethics of the research process for this study were observed diligently, especially in consideration of sampling working women in developing regions, with the absence of structural and legal support [50]. All participants were informed and assured by attaching a cover letter to the questionnaire, describing the objectives of the research and ensuring confidentiality and anonymity. Informed consent was taken from the participants. Private rooms were requested in nursing schools where all surveys were filled and respondents were able to complete the survey in privacy and also to ask any questions related to the survey. Face and content validity of the questionnaire was confirmed through consultation and discussion with senior researchers, nurse supervisors and nurse ward heads. Cronbach’s alpha was used to check the internal consistency and reliability of the items in the instrument [51, 52]. Internal validity was ensured by using a simple random sample so that each participant had an equal chance of selection. Construct validity was assured by using PCA.

The Kaiser-Meyer Olkin test and the Bartlett test were both satisfactory and supportive to conducting a factor analysis. Six factors were extracted by PCA, including all the six subscales for organizational culture represented in the PES-NWI. Results showed 70.8 % of the variance of the construct being studied (i.e. organizational culture). Communalities were extracted (Table 2), and each item explains at least 50 % of the variance of the study construct.

The nurse respondents average composite mean score for organizational culture was ranked favorably at 2.38 (SD = 0.616). The following three areas were ranked unfavorably by nurses (Table 3): (i) error reporting culture (Mean score = 2.62; SD = 0.500), (ii) staffing and resource adequacy (Mean score = 2.56; SD = 0.901), and (iii) nurse foundations for quality of care (Mean score = 2.59; SD = 0.630). The following subscales of organizational culture have been ranked favorably by nurse respondents: (i) nurse participation in hospital affairs (Mean score = 2.34; SD = 0.726), (ii) nurse participation in governance (Mean score = 2.33; SD = 0.781), and (iii) nurse-coworkers relations (Mean score = 2.26; SD = 0.704).

Of the 309 respondents (Table 4), 202 were from Hospital A (65.4 %) and 107 were from Hospital B (34.6 %). About half of the nurse respondents was in the age group of 20–29 years (n = 161, 52.1 %), 87 belonged to the age group of 30–39 years (28.2 %) and 19.7 % (n = 61) were 40 years of age and above. The average age was 30.5 years (SD = 9.85). The majority was currently married (n = 173, 55.9 %); 130 respondents (42.1 %) were unmarried and 6 respondents (1.9 %) were divorced, widowed or separated. Nearly all respondents belonged to the province of Punjab (n = 301, 97.4 %), 6 were from Sindh (1.9 %) and one each (0.3 %) was from Baluchistan and Khyber Pakhtun Khwan, respectively. Most of the nurse respondents were Muslims (n = 231, 74.7 %), 76 were Christians (24.6 %) and one respondent each (0.3 %) was Hindu and Ahmedi. More than half of the respondents (n = 157, 50.8 %) had no children, 63 had one or two children (20.4 %) and 89 had three children or more (28.8 %). A total of 56 nurses earned between PKR 5,000–19,999 (18.1 %), 154 nurses earned between PKR 20,000–39,999 (49.8 %) and 99 nurses earned more than PKR 40,000 (32.0 %). The average income per month of the nurses in the sample was PKR 33,754. With regard to the place of residence, the majority (n = 212, 68.6 %) of nurse respondents lived in private accommodations, 53 (17.2 %) of nurse students lived in college hostel and 44 (14.2 %) lived in the hospital residence colony.

A total of 174 (56.3 %) nurses had a highest nursing degree of diploma (Table 5), 120 (38.8 %) had earned a BSc in Nursing and 15 (4.8 %) had an MSc in Nursing. Majority of the nurses (n = 205, 66.3 %) had a designation of staff nurse, 51 were student nurses (16.5 %), 40 nurses were ward heads (12.9 %), 11 were nurse instructors (4.2 %) and two were nurse supervisors (0.6 %). A little more than half of the nurse respondents had a permanent position (n = 176, 57.0 %). Nearly all the nurse respondents were full-time employees (n = 94.8 %). Most of the nurse respondents belonged to the 16 grade government scale (n = 256, 82.8 %). 73.8 % (n = 228) of the nurses were not working at a private clinic after duty hours at the hospital, whereas 81 (26.2 %) were working at a private clinic. A total of 59 nurses reported having to work additional hours as night duty (19.1 %), 128 reported having to work additional hours in the day (41.4 %) and 124 had to work additional hours in the evening (40.1 %).

Composite scores for organizational culture, the six subscales of organizational culture and error reporting were calculated (Table 6). The normality assumption was evaluated and scores of all subscales of organizational culture followed the normal distribution. The correlations, using Pearson correlation, between all study variables significantly correlated with values of above 0.3. Correlation coefficients also showed that the variables have a positive relationship and move together in a linear fashion.

The contingency results for regression have been shown in Table 7. Results for bivariate logistic regression (Table 8) show that the composite organizational culture variable and the subscales of organizational culture all have high odd ratios with favorable culture of error reporting. When organizational culture is favorable, nurses perceived higher odds of error reporting (OR: 2.43, 95 % CI: 1.51–3.92). When nurse participation in governance (OR: 1.83, 95 % CI: 1.16–2.87) and nurse participation in hospital affairs (OR: 2.96, 95 % CI: 1.85–4.70) are favorable, there is a higher odds of error reporting. Similarly, when nurse manager ability, leadership and support (OR: 1.56, 95 % CI: 0.98–2.48), nurse foundations for quality of care (OR: 3.12, 95 % CI: 1.96–4.98), nurse staffing and resource adequacy (OR: 7.83, 95 % CI: 4.64–13.22) and nurse–coworker relations (OR: 6.13, 95 % CI: 3.62–10.37) were favorable, the odds of error reporting was high. The results also show that nurses above the age of 30 years had extremely higher odds of reporting errors (OR: 13.73, 95 % CI: 7.91–23.86). Married nurses (OR: 5.54, 95 % CI: 3.39–9.05) and nurses earning an income of above PKR 40,000 (OR: 2.55, 95 % CI: 1.54–4.21) had higher odds of reporting errors. Also, nurses on a permanent contract (OR: 6.98, 95 % CI: 4.21–11.57) were more likely to report errors.

Multivariable logistic regression was performed to calculate the adjusted odds ratio (AOR), holding income, education and age (as a continuous variable) as constants (Table 8). Results were highly significant for the odds ratios between error reporting and organizational culture and its subscales. When organizational culture (AOR: 3.58, 95 % CI: 1.93–6.63), nurse participation in governance (AOR: 3.33, 95 % CI: 1.87–5.95), nurse participation in hospital affairs (AOR: 5.08, 95 % CI: 2.69–9.57), nurse manager ability, leadership and support (AOR: 2.61, 95 % CI: 1.40–4.84), nurse foundations of quality of care (AOR: 4.83, 95 % CI: 2.59–9.02), nurse staffing and resource adequacy (AOR: 7.86, 95 % CI: 4.18–14.75) and nurse coworker relations (AOR: 5.58, 95 % CI: 2.97–10.50) were all favorable, the odds of error reporting were significantly higher. Also, married nurses (AOR: 1.33, 95 % CI: 0.17–0.63) and nurses with a permanent contract (AOR: 1.29, 95 % CI: 0.14–0.599) had higher odds of reporting errors.

Our sample respondents ranked the error reporting culture in their hospitals as unfavorable. This has significant implications for other public sector hospitals in the region, since previous research confirms that no formal error tracking systems exist in the Pakistan healthcare setup and also that education in ethics observance and administrative policy measures for promotion of ethical cultures are absent [53]. Currently, there are no all-encompassing state medical laws in the country for the safeguard of either patients or medical and nurse practitioners. There may be several reasons for the absence of an error reporting culture in the hospital settings of Pakistan, as perceived by female nurses in this study. First, senior doctors and physicians have an elevated status and elite labels, which supports bullying and blame-shifting in the hospital setting against junior medical practitioners and nurses [54]. Second, the healthcare sector of the region has a blame culture with punitive action taken against individuals, which prevents individuals from error sharing [55]. Third, nursing is a feminized profession in the region, with female nurses reluctant to report errors due to male dominated and patriarchal work environments and the fear of having to face workplace violence and retribution [19].

Study results also show that hospital staffing and resource adequacy is perceived by nurse respondents as unsatisfactory. Severe shortages in resources and staffing in the public sector hospitals of Pakistan critically undermine efforts of medical practitioners to ensure patient safety [14]. Other research from the region confirms that corruption in the health care sector, with public sector hospitals commonly devoid of basic and life-saving medicines and medical equipment, is responsible for both shortages and high rates of mortality [56]. Of the total national gross domestic product, only 2.9 % is spent on health care and only 1.2 % is allocated to the public sector [57]. Although more than 70 % of the service provision in health care is provided by the private sector in the region, it is estimated that 74 % of the population of Pakistan avail public health care services due to lack of funds [57]. In addition, public sector staffing is lacking due to inadequate budget allocations for hiring and compensation of medical and nurse practitioners. This has led to low enrollment and high rates of immigration [58].

The nurse professional is an integral member of the health care sector who is responsible directly for patient safety, the efficiency of the health care organization and the overall wellbeing of the population [59]. The results of this study, however, highlight that nurse foundations for quality of care in the hospital administration are ranked as unfavorable by respondents. Other research also suggests that Pakistani hospitals are dominated by medical care plans, with little attention to nursing care plans [60]. This may be because nursing is a feminized profession in the region, and male dominated medial administrations give minimal emphasis to nursing care plans for patient care delivery and instead give prominence to medical care plans [61]. Nursing is perceived in patriarchal regions as a care provision, restricted to cleaning, washing and execution of orders passed by doctors and physicians [62]. Non-nurse medical practitioners, medical administrators and patients do not recognize that nurses have medical training and are aware of patient’s medical needs. In this way nursing care plans are not given precedence.

An unexpected finding was that nurses scored other subscales of organizational culture, such as nurse participation in governance, nurse participation in hospital affairs, nurse manger ability, leadership and support, and nurse coworker relations as favorable. Previous studies in the region indicated that nurse autonomy, participation in hospital policy-making and teamwork are extremely unfavorable [12‐14, 17]. It may be that nurse perceptions of favorable organizational culture are highly dependent on nurse education, training and exposure. For example, nurses in the sample may not be comparing their work environment to hospitals in developing nations. They could be strongly influenced by cultures in their wider community. Their perception of the organizational culture at their workplace may be better than their domestic and home environments [63]. There is also the disadvantage of attempting to collect survey data about sensitive topics in blame cultures and patriarchal communities, which may influence female respondents in indicating anything negative about governance and management.

This study furthermore confirms, through multivariable regression results, that for an improved error reporting culture, organizational culture and all its six categories (nurse manager ability, leadership and support, nurse participation in hospital affairs, nurse participation in governance, nurse foundations of quality care, nurse-coworkers relations, and nurse staffing and resource adequacy) need to be favorable. Regression results support findings from other international research [43, 64, 65]. Also, our findings show that error reporting is more common amongst nurses who are married and on a permanent contract. Other studies show that married nurses are more likely to participate in studies about errors [66], and this may be because married women are more secure in facing the consequences of error reporting, like blame, shame, retribution and job loss, due to the safety of a dual-income earning household. Furthermore, nurses on permanent contract may be more likely to report errors because of the difficulty in having their state contracts or jobs rescinded, due to extremely slow bureaucratic processing by the government [56].

There are several limitations of this study. The size of the sample is relatively small and excludes public sector tertiary care hospitals across other cities and rural areas of Pakistan. The low response rate of 34.8 % was due to the lack of time available of busy nurses and also the unwillingness of nurses to participate in what was considered a sensitive topic in a male-dominated work organization. In addition, the responses of nurses are guided by their perceptions, which are influenced by their level of education, on-going training and exposure to magnet hospitals. Also, because of the small sample, findings cannot be generalized. Despite the limitations, this study has significant strengths. It is the only research from Pakistan assessing the relationships between nurse perceptions of organizational culture and the culture of error reporting in public sector hospitals. We hope that our study’s findings will have wider macro implications, as improved patient safety is known to help improve overall public health and reduce health costs for the national economy. Additionally, the findings highlight the critical shortages in staffing and resources and the inadequacy of nursing care plans for patient safety culture in the healthcare sector of the region.